benchmark result I don't trust: anyone testing this?

[benchplease]

Did a bit of testing and thought I'd post the result here.

After a few runs on dual 3090, I ended up with decent performance, but the setup only really made sense once I adjusted around router policy.

Nothing scientific here, just practical notes from a real setup.

[northgrid33]

I tested this tonight. My result was mixed: first run looked bad, second run looked normal.

[homelabhenry]

My first guess would be this: If you're on llama.cpp, I'd confirm the version first and then reduce variables until the behavior is obvious. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[cacheline]

A couple questions before I guess too hard. Are you testing this on dual 3090 or something else?

[northhall88]

I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with FLUX and compare from there.

[quiethouse]

I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with FLUX and compare from there.

Maybe, but I had the opposite result. I've seen people blame VRAM headroom when the real issue was much more mundane. Worth testing the simpler explanation before adopting the more interesting one.

[benchplease]

I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with FLUX and compare from there.

worth checking the driver first.

[cacheline]

What exact setup are you running? What model, what runtime, and roughly what context size?

[northgrid33]

A couple things I'd check first. If you're on vLLM, I'd confirm the version first and then reduce variables until the behavior is obvious.

[nanobrook]

I'd approach it in a boring order. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at tail latency. Not glamorous advice, but repeatable baselines save a lot of time.